As you know, every week we are featuring a set of predictions from a model developed by the Michigan Soccer Analytics Society; Mitchel Green created the model, and Charlie Tuley has adapted it for La Liga. (You can see last week’s predictions here). A lot of you both on the site and on twitter have expressed a lot of interest in this, and Charlie is here to explain to people like me, who aren’t analytics people, how the model works, what it can capture, and what it can’t.
Zach Hicks is like a cat lapping up cream when it comes to analytics, and he and Charlie discuss the concepts behind the model, what inputs go into it, how the model ‘learns’ from previous matches, what potential changes could be made (and which ones wouldn’t really achieve much) and a lot besides. We had a great time and I am sure we’ll have Charlie back on again to see how the model is working out as we move through our season. Enjoy!